Models of computation for embedded system design

In the near future, most objects of common use will contain electronics to augment their functionality, performance, and safety. Hence, time-to-market, safety, low-cost, and reliability will have to be addressed by any system design methodology. A fundamental aspect of system design is the specification process. We advocate using an unambiguous formalism to represent design specifications and design choices. This facilitates tremendously efficiency of specification, formal verification, and correct design refinement, optimization, and implementation. This formalism is often called model of computation. There are several models of computation that have been used, but there is a lack of consensus among researchers and practitioners on the “right” models to use. To the best of our knowledge, there has also been little effort in trying to compare rigorously these models of computation. In this paper, we review current models of computation and compare them within a framework that has been recently proposed. This analysis demonstrates both the need for heterogeneity to capture the richness of the application domains, and the need for unification for optimization and verification purposes. We describe in detail our CFSM model of computation, illustrating its suitability for design of reactive embedded systems and we conclude with some general considerations about the use of models of computations in future design systems.